Then there’s Android Wear, a wearable-tuned variant of Google’s Android operating system, found in multiple manufacturers’ products … including LG’s G Watch series and the non-LTE version of LG’s Watch Urbane, the Samsung Galaxy Gear Live, and Sony’s latest SmartWatch generation. As you can see, some companies are hedging their bets. The Android Wear supplier ecosystem also includes plenty of other familiar company names, among them Motorola, formerly a division of Google and now owned by Lenovo.
Although Motorola was publicised as a launch partner when Google unveiled Android Wear in March 2014, the Moto 360 smart watch didn’t begin shipping until September of that same year (LG’s first-generation G Watch and Samsung’s Gear Live had entered production in June, coincident with the Google I/O developer conference). The Moto 360, as its name implies, is unique among Android Wear-based smart watches for its circular face (at least as I write these words … rumours suggest that it’ll soon be joined by products from others). This particular form factor decision is cosmetically preferable, at least to me, but it also compelled some design compromises.
My motivation in tackling this particular teardown is to showcase these tradeoffs as well as, more generally, to highlight what’s inside the Moto 360’s sleek case. Given that the product’s been available since early last autumn, I suspected it had already been dissected elsewhere, and Google informed me that indeed it had … specifically by my long-time friends at iFixit. My perusal of their teardown report’s travails and other details, along with a subsequent conversation with company CEO Kyle Wiens, was informative on a number of fronts.
First off, I quickly realised that without the necessary specialised tools, there was a near-100% certainty that disassembly attempts would render my smart watch no longer usable. Even with the appropriate gear in hand, the iFixit gang had snapped the rear cover of their product sample in two. To wit, I frankly feared that my fumbling might even mar the hardware so badly that subsequent teardown analysis would be significantly hampered or completely precluded. And of course, the Moto 360’s functional demise would preclude ongoing hands-on coverage.
Wiens instead offered to share any and all photos from his company’s Moto 360 project with me, echoing multiple past partnerships between iFixit and EDN Magazine (me, specifically). I gratefully and enthusiastically accepted. Although the Moto 360 hardware is unchanged from last September, the firmware’s gone through one update so far. Tangibly longer battery life was a notable outcome of the upgrade, along with other enhancements. A pending (as I write these words) additional update will reportedly make additional improvements:
The ability to keep the watch and paired tablet or smartphone connected even if they’re not in Bluetooth proximity, as long as they’re both connected to the "cloud." In the smart watch case, this cloud connectivity is accomplished by activating (with unknown battery life impacts) the Moto 360’s Wi-Fi subsystem; the watch and paired mobile device don’t even need to be on the same LAN subnet.
Touchscreen user interface improvements that, among other things, speed access to applications installed on the watch.
The ability to more robustly control (beyond today’s hit-and-miss LCD backlight activation) the watch’s UI (including scrolling through notifications) via wrist rotations, etc. in conjunction with the Moto 360’s built-in gyro and accelerometer.
Rumour also has it that the Moto 360 and other Android Wear smart watches will sooner-or-later be compatible with not only Android-based smartphones and tablets but also Apple iOS-based mobile devices. Time will tell whether or not this particular potential enhancement ends up happening. Regardless, it’s clear that Motorola and Google’s software enhancement exercises, which also include Android apps (and regular updates to same), are effective in keeping the ageing hardware working effectively. About that hardware … let’s have a look, shall we?
After removing the Moto 360’s plastic back cover, we can get a better view of the built-in heart rate sensor. Here’s a description of how it works, ironically from an Apple Watch support document, courtesy of John Gruber’s Daring Fireball site (and Gizmodo has a nice photo of the feature in operation):
The heart rate sensor in Apple Watch uses what is known as photoplethysmography. This technology, while difficult to pronounce, is based on a very simple fact: Blood is red because it reflects red light and absorbs green light. Apple Watch uses green LED lights paired with light-sensitive photodiodes to detect the amount of blood flowing through your wrist at any given moment. When your heart beats, the blood flow in your wrist — and the green light absorption — is greater. Between beats, it’s less. By flashing its LED lights hundreds of times per second, Apple Watch can calculate the number of times the heart beats each minute — your heart rate.
The photoplethysmography sensor can be used for more than just pulse rate monitoring purposes; it’s theoretically capable of also acting as a pulse oximeter for assessing your oxygen saturation level. However, as far as I know, these enhanced SpO2 capabilities are not (yet) enabled, either for the Moto 360 and other Android Wear-based watches or, for that matter, the Apple Watch. Underneath the FCC sticker (ID IHDT6QC1) shown in the upper of the two photos, by the way, is a five-terminal spring-contact pad connector with unknown function; production line programming and testing, perhaps? Or an expansion bus for as-yet unnanounced "smart" wrist bands, analogous to those planned for the Pebble Time line? The Apple Watch is similarly adorned, come to think of it.
Here’s what it looks like after you succeed in separating the multi-layer interior "sandwich" (bottom) from the front case assembly containing the display (top). Note the ribbon cable connecting them. Note, too, the lime green water resistance-augmenting rubber O-ring around the assemblage. The Moto 360 is rated IP67, which translates to complete protection from dust ingress, along with protection against liquid immersion up to a metre deep for 30 minutes.
Next is the system board-plus-battery sandwich (right), separated from the rear housing (left). Somewhat surprisingly (at least to me), particularly considering the pervasive concern about smart watch battery life regardless of manufacturer or model, Motorola went with a conventional four-sided lithium ion polymer battery (3.8V, 300 mAh minimum/320 mAh typical, ~1.1 Wh), versus a higher capacity (or alternatively, thinner for equivalent capacity) round battery that would more completely fill the rear housing cavity. Admittedly, going the conventional-form-factor route saved Motorola bill-of-materials cost (a theme we’ll again encounter soon), and the company seems to have found other ways of using the available rear housing space. Still … I’m a bit surprised.
Let’s now look more closely at the rear housing. We already knew that it encompasses the photoplethysmography sensor. But as you may also already be aware, the Moto 360 is a wireless-only charged device, specifically via the Qi inductive power transfer standard. Peel away the sticker which shields the circuit board from any misbehaving energy, and the wireless receiver coil underneath is exposed for your perusal.
Here’s its transmitter coil companion, in the charger dock. The Moto 360 fits snugly against it and it’s micro-USB interface-equipped. Motorola included a wall wart to power it, but you can always alternatively make do with a micro-USB to USB cable in combination with an USB-output wall wart, multi-port charger, or sufficiently powerful computer USB port.
After separating the battery from the system board, it’s finally time to examine the latter in greater detail. In the upper right corner is Texas Instruments’ TMS320C5545 fixed-point DSP, which I suspect is present to handle the smart watch’s "Ok Google" speech recognition capabilities. To its left is a mysterious IC labeled "WL18G/31/46C1VRI $N," which as it turns out is a Texas Instruments-sourced wireless transceiver module, handling Wi-Fi, Bluetooth, and Bluetooth Low Energy protocols.
Let’s now look at the two large ICs below the TMS320C5545-plus-WL18G combo, and extending down the center of the board. The upper, labeled as 2SB28 D9QRM, is actually a Micron Technology MT46H128M32L2KQ-5 IT 4 Gbit mobile LPDDR SDRAM. The lower is a Toshiba THGBMAG5A1JBAIT 32 Gbit NAND flash memory, with an e-MMC interface, implementing the smart watch’s 4 GBytes of resident OS, application, and data storage.
Next, let’s peruse the ICs along the right side of the PCB. Next to the Micron SDRAM is Solomon Systech SSD2848K1 (PDF) display controller, implementing the MIPI interface protocol in driving the LCD. Below it is Atmel’s MXT112S capacitive touchscreen controller, supporting the LCD’s touch interface facilities. And below it (and directly next to the Toshiba flash memory chip) is Texas Instruments’ AFE4490, which acts as an analogue front-end companion to the previously mentioned pulse oximeter sensor.
Finally, let’s peruse the PCB left-side ICs. In the upper corner of the Micron SDRAM is a Texas Instruments 1211A1 USB 2.0 PHY transceiver. Its presence is admittedly a bit baffling to me, since the Moto 360 offers no to-outside-world USB connectivity, but perhaps one or more of the smart watch’s internal subsystems requires a USB interface. Below it and directly to the side of the MT46H128M32L2KQ-5 IT is Texas Instruments’ TPS659120 (PDF) power management unit. And to the side of the Toshiba flash memory is Texas Instruments BQ51051B, which interfaces with the Qi wireless power receiver for battery-charge management purposes. If you’ve been keeping a running tally, you’ve undoubtedly already noticed what a bag o’chips design win the Moto 360 represents for TI!
Over on the far left of the PCB…
Over on the far left of the PCB…are two other ICs, whose function begs for more explanation. The shiny one at top is Wolfson Microelectronics’ (now Cirrus Logic’s) WM7132 MEMS microphone, with a bottom-side sound input port location (whose ambient-air access scheme will become clear shortly). Its companion is the WM7121, with a topside port that you can see if you look closely. The two microphones work in a tandem "array" arrangement (that I’ve been writing about for years), which both allows the DSP connected to them to "localise" a desired sound source and, in the process, suppress spurious ambient noise.
Their shared access to the outside world consists of a small hole in the left side of the Moto 360 (opposite the watch’s sole right-side button), shown above and which I admittedly didn’t even notice in my unit until after I’d learned where the microphones inside could be found. I’m guessing that a membrane behind the hole allows environmental sound vibrations, but not moisture or dust, to pass through. But speaking of membranes, note that the Moto 360 is silent; there’s no speaker inside, only a tiny motor for vibration, which along with the display constitutes the sole means of user communications.
Now, let’s take a look at the PCB backside. It’s mostly bare; remember that the battery is normally attached to it. But it’s not completely uninteresting. First, remember the just-mentioned WM7132 bottom-port microphone? Look closely at the PCB’s right edge in the above photo and you’ll see a small hole which, along with the hole in the side of the Moto 360, gives the MEMS microphone access to sounds coming from the outside world. At the bottom of the PCB are five contacts, which correspond (thanks to spring-fed intermediaries) to the previously mentioned mysterious five-contact port underneath the smart watch’s FCC certification sticker. And in the PCB’s upper left corner you’ll find the Moto 360’s combo six-axis accelerometer/gyro, an InvenSense MPU-6050 MEMS motion tracking device, to be precise.
One important IC is absent from the so-far discussion. Without reading on, do you know what it is? I haven’t yet mentioned the primary system CPU or, for that matter, the GPU that drives the display. They’re all together in one application-tailored SoC, Texas Instruments OMAP3630, labelled "X3630ACBP," located underneath the Micron SDRAM and shown in the photo below:
The OMAP3630 integrates (among other things) a single ARM Cortex-A8 processor core running at up to 1 GHz, along with a PowerVR SGX 530 graphics core running at up to 200 MHz and a C64X DSP core running at up to 800 MHz. Motorola has caught a fair bit of heat in the technical press for the decision to use a 2010-era SoC, much of it IMHO unjustified. I don’t have insight into what clock speeds the Moto 360 is actually using for the various cores in the chip; what I can say, however, is that the smart watch’s responsiveness is consistently snappy, with no display stutter or other discerned lag. If the OMAP3630 is sufficient for the task at hand, why bother using something even higher in performance (and potentially higher priced), right?
With that overriding observation said, I’ll offer a few critiques. For one thing, the OMAP3630 was originally fabricated on a 45 nm process. TI may have subsequently transitioned the design to a more modern and shrunk lithography node, but if it hasn’t, the semiconductor foundation for the application SoC is probably more power-hungry than it needs to be. The other advantage of a smaller-dimension process is that it delivers a greater transistor budget for a given-sized sliver of silicon. By using a more advanced ARM Cortex-A9 (or even newer) processor core, for example, Motorola might have gained access to the NEON SIMD accelerator.
More transistors also might have allowed for the inclusion of a more advanced dedicated DSP core. And either or both advancements might have precluded the need for the separate TMS320C5545 DSP in the design. With that said, I can see at least one other advantage of the distributed processing approach, if my guess is correct that the discrete DSP is doing speech recognition. The TMS320C5545 can remain awake, monitoring for "Ok Google" user utterances while the rest of the smart watch’s processing resources are slumbering, thereby minimising overall power consumption.
With the PCB detailed, let’s now turn our attention to the display. Above are two views of it and its companion ribbon cable and other circuitry. While other smart watches make use of E ink (Pebble) or OLED (Apple Watch, other Android Wear-based watches) display technologies, Motorola went with a circular LCD. I realise that Motorola’s LCD choice is advantageous in the ability to view the display in bright ambient light versus with the more common (but washed-out) OLED alternative, as well as delivering a rich luminance and chrominance palate versus an E ink monochrome or "colour" display alternative. But the fundamental downside of the LCD, the battery-draining backlight, can’t be ignored, either.
My Moto 360 requires recharge at least every 1.5 days, if not sooner; every few hours in its optional Ambient Screen mode, which keeps the backlight at-minimum dimly illuminated at all times. I suspect that Motorola’s primary LCD motivation was cost; circular screens aren’t mainstream (translation: cheap) regardless of their technology foundation, but a mature LCD basis was probably still less costly than the upstart OLED alternative would have been. Still, at the end of the day, I can’t help but wonder if a circular (and no-backlight-required) OLED might have been the better choice, especially considering that limited display lifetime isn’t as much an issue with a watch as it is, say, with a TV. And I’m even belatedly warming to Qualcomm’s longstanding Mirasol display technology advocacy.
One other display-related topic bears mentioning. Some of you may have heard about the Moto 360 LCD’s "flat tyre" characteristic, which you can clearly see in the upper of the two photos above. What you’re looking at is not some flaw in the display itself; rather, this is where Motorola chose to locate the ambient light sensor. As the second photo above suggests, the "flat tire" is only obvious when a light-colour watch face is in use; with a mostly-black face such as the "Minimal" option shown (which I personally also use), it’s only obvious if you consciously look for it. And with both a black face and black case, as with the product photo at the beginning of this writeup, it’s essentially a non-issue.
Finally, let’s circle back to something I discussed at the beginning of the writeup, that being the pending firmware update that will awaken the Moto 360’s now-dormant Wi-Fi facilities. As you’ve already seen, the Texas Instruments WL18G wireless transceiver module inside the watch supports them. But where’s the corresponding antenna, either for Wi-Fi or now-active Bluetooth? Embedding one or both antennas within the metal ring encasing the watch is one possibility; AnandTech alludes to this in its product-launch coverage:
We see custom antennas that are in the metal housing but don’t require any antenna lines. Unfortunately, there was no real disclosure on how this worked so it was hard to say how they pulled this off but there are noticeable patterns on the inside of the metal casing. New RF techniques were also used to make custom metal wristbands that don’t interfere with the antennas of the watch itself.
But I haven’t seen any sort of physical cable connection between the metal case and PCB to bolster this particular supposition. One other possibility is that one or both antennas might instead be embedded within the PCB. Texas Instruments’ WL18G datasheet (PDF) even documents such an approach in some detail.